Efficient Multi-Core Computations in Computational Statistics and Econometrics

Abstract

The social researchers use computationally intensive statistical and econometric methods for data analysis. One way for accelerating these computations is to use the parallel computing with multi-core platforms. In this paper we parallelize some representative computational kernels from statistics and econometrics on multi-core platform using the programming libraries such as Pthreads, OpenMP, Intel Cilk++, Intel TBB, Intel ArBB, SWARM and Fast Flow. Specifically, these kernels are multivariate descriptive statistics (such as multivariate mean and multivariate covariance) and kernel density estimation (univariate and multivariate). The purpose of this paper is to present an extensive quantitative and qualitative study of the multi-core programming models for parallel statistical and econometric computations. Finally, based on this study we conclude that the Intel ArBB and the SWARM programming environments are more efficient for implementing statistical computations of large and small scale, respectively. The reason for which these models are efficient because they give good performance and simplicity of programming.